Differences in the Binding Affinity of an HIV-1 V2 Apex-Specific Antibody for the SIVsmm/mac Envelope GlycoproteinUncouple Antibody-Dependent Cellular Cytotoxicity fromNeutralization

Benjamin von Bredow,a* Raiees Andrabi,b,c Michael Grunst,a Andres G. Grandea III,a Khoa Le,b,c Ge Song,b,c

Zachary T. Berndsen,c,d Katelyn Porter,b,c Jesper Pallesen,c,d* Andrew B. Ward,c,d Dennis R. Burton,b,c,e David T. Evansa,f

aDepartment of Pathology and Laboratory Medicine, University of Wisconsin, Madison, Wisconsin, USAbDepartment of Immunology and Microbiology, The Scripps Research Institute, La Jolla, California, USAcInternational AIDS Vaccine Initiative Neutralizing Antibody Center, the Collaboration for AIDS Vaccine Discovery (CAVD) and Center for HIV/AIDS Vaccine Immunology-Immunogen Discovery (CHAVI-ID), The Scripps Research Institute, La Jolla, California, USA

dDepartment of Integrative Structural and Computational Biology, The Scripps Research Institute, La Jolla, California, USAeRagon Institute of MGH, MIT and Harvard, Boston, Massachusetts, USAfWisconsin National Primate Research Center, University of Wisconsin, Madison, Wisconsin, USA

ABSTRACT As a consequence of their independent evolutionary origins in apes andOld World monkeys, human immunodeficiency virus type 1 (HIV-1) and simian im-munodeficiency viruses of the SIVsmm/mac lineage express phylogenetically and anti-genically distinct envelope glycoproteins. Thus, HIV-1 Env-specific antibodies do nottypically cross-react with the Env proteins of SIVsmm/mac isolates. Here we show thatPGT145, a broadly neutralizing antibody to a quaternary epitope at the V2 apex ofHIV-1 Env, directs the lysis of SIVsmm/mac-infected cells by antibody-dependent cellu-lar cytotoxicity (ADCC) but does not neutralize SIVsmm/mac infectivity. Amino acidsubstitutions in the V2 loop of SIVmac239 corresponding to the epitope for PGT145in HIV-1 Env modulate sensitivity to this antibody. Whereas a substitution in a con-served N-linked glycosylation site (N171Q) eliminates sensitivity to ADCC, a lysine-to-serine substitution in this region (K180S) increases ADCC and renders the virus sus-ceptible to neutralization. These differences in function correlate with an increase inthe affinity of PGT145 binding to Env on the surface of virus-infected cells and tosoluble Env trimers. To our knowledge, this represents the first instance of an HIV-1Env-specific antibody that cross-reacts with SIVsmm/mac Env and illustrates how differ-ences in antibody binding affinity for Env can differentiate sensitivity to ADCC fromneutralization.

IMPORTANCE Here we show that PGT145, a potent broadly neutralizing antibody toHIV-1, directs the lysis of SIV-infected cells by antibody-dependent cellular cytotoxic-ity but does not neutralize SIV infectivity. This represents the first instance of cross-reactivity of an HIV-1 Env-specific antibody with SIVsmm/mac Env and reveals that an-tibody binding affinity can differentiate sensitivity to ADCC from neutralization.

KEYWORDS ADCC, antibody function, human immunodeficiency virus, neutralizingantibodies, simian immunodeficiency virus

HIV-1 and SIVsmm/mac evolved independently in apes and Old World monkeys andbelong to phylogenetically distinct lineages of primate lentiviruses (1). Whereas

HIV-1 is a result of the cross-species transmission of SIVcpz from chimpanzees intohumans (2, 3), SIVmac, which is widely used as a model for HIV-1 infection in nonhumanprimates, is a result of the transmission of SIVsmm from sooty mangabeys to Asian

Citation von Bredow B, Andrabi R, Grunst M,Grandea AG, III, Le K, Song G, Berndsen ZT,Porter K, Pallesen J, Ward AB, Burton DR, EvansDT. 2019. Differences in the binding affinity ofan HIV-1 V2 apex-specific antibody for theSIVsmm/mac envelope glycoprotein uncoupleantibody-dependent cellular cytotoxicity fromneutralization. mBio 10:e01255-19. https://doi.org/10.1128/mBio.01255-19.

Invited Editor Manish Sagar, BostonUniversity

Editor Jaisri R. Lingappa, University ofWashington

Copyright © 2019 von Bredow et al. This is anopen-access article distributed under the termsof the Creative Commons Attribution 4.0International license.

Address correspondence to David T. Evans,dtevans2@wisc.edu.

* Present address: Benjamin von Bredow,Department of Pathology and LaboratoryMedicine, Penn State College of Medicine,Hershey, Pennsylvania, USA; Jesper Pallesen,Department of Molecular and CellularBiochemistry, Indiana University, Bloomington,Indiana, USA.

Received 17 May 2019Accepted 29 May 2019Published 2 July 2019

RESEARCH ARTICLEHost-Microbe Biology

crossm

July/August 2019 Volume 10 Issue 4 e01255-19 ® mbio.asm.org 1

on May 25, 2020 by guest

http://mbio.asm

.org/D

ownloaded from

species of macaques (4–6). As a reflection of their differing evolutionary histories, HIV-1and SIVsmm/mac are antigenically distinct, and antibodies to the gene products of onevirus typically do not cross-react with those of the other (7, 8). This is especially true forEnv, which is rapidly evolving under the selective pressure of host antibody responses(9, 10). Thus, studies to assess protection by HIV-1-specific antibodies in nonhumanprimates depend on the use of chimeric simian-human immunodeficiency viruses(SHIVs) expressing HIV-1 Env proteins (11).

Recent studies by our group and others have shown that the antibody-dependentcellular cytotoxicity (ADCC) of antibodies to the HIV-1 envelope glycoprotein generallycorrelates with their ability to neutralize viral infectivity (12–14). Nevertheless, instancesof neutralization in the absence of ADCC were observed, revealing differences in Envepitopes exposed on the surface of virions versus infected cells (12). Instances of ADCCin the absence of neutralization were also observed for a few antibodies, but onlyagainst lab-adapted HIV-1NL4-3 (12). For primary HIV-1 isolates, every antibody withdetectable ADCC activity neutralized viral infectivity, suggesting that the epitopesexposed on the surface of cells infected with these viruses are also accessible onfunctional Env trimers of virions (12).

In the present study, we show that PGT145, which recognizes a quaternary proteoglycanepitope at the V2 apex of HIV-1 gp120 (15–17), is able to opsonize SIVsmm/mac-infected cells for NK cell lysis. The ADCC activity of PGT145 correlates with Env stainingon SIVsmm/mac-infected cells; however, this antibody does not neutralize SIVsmm/mac

infectivity. To our knowledge, this represents the first instance of an HIV-1 Env-specificantibody cross-reacting with the Env proteins of SIVsmm/mac and reveals that differencesin the affinity of antibody binding to Env can determine sensitivity to ADCC versusneutralization.

RESULTSPGT145 targets HIV- and SIV-infected cells for antibody-dependent cellular

cytotoxicity. In the course of testing broadly neutralizing antibodies for ADCC activityagainst HIV-1-infected cells, we found that the V2 apex-specific antibody PGT145mediated ADCC, not only against HIV-infected cells, but also against cells infected withSIVmac239 (Fig. 1A). To assess the breadth of this cross-reactivity, ADCC was measuredagainst cells infected with three different SIV isolates, including SIVmac239, aneutralization-resistant virus that predominantly infects T cells (18, 19); SIVmac316,which is a macrophage-tropic isolate that is particularly sensitive to neutralizingantibodies (20, 21); and SIVsmE543-3, which is an independently derived SIV isolate thatexpresses a distinct, neutralization-resistant envelope glycoprotein (22). In addition todirecting the killing of SIVmac239-infected cells, PGT145 mediated NK cell lysis of cells

FIG 1 PGT145 mediates ADCC against HIV-, SHIV-, and SIV-infected cells. CEM.NKR-CCR5-sLTR-Luc cells infected withthe indicated strains of HIV-1, SHIV, and SIV (A) and SIV alone (B) were incubated with an NK cell line (KHYG-1 cells)expressing human CD16 in the presence of the indicated concentrations of PGT145. ADCC activity was measuredas the dose-dependent loss of luciferase activity in % RLU, compared to control wells containing no antibody andNK cells with either infected (maximal) or uninfected (background) target cells. The dotted line representshalf-maximal lysis of infected cells, and the error bars represent the standard deviation of the mean from triplicatewells.

von Bredow et al. ®

July/August 2019 Volume 10 Issue 4 e01255-19 mbio.asm.org 2

on May 25, 2020 by guest

http://mbio.asm

.org/D

ownloaded from

infected with SIVmac316 and SIVsmE543-3 (Fig. 1B), indicating that these responses toSIV-infected cells were not strain specific.

PGT145 binds to cells infected with HIV, SHIV, and SIV. To further investigate theunexpected lysis of SIV-infected cells by PGT145, binding of this antibody to HIV-1-,SHIVAD8-EO-, and SIV-infected cells was assessed by flow cytometry. Env staining on thesurface of infected cells corresponded with susceptibility to ADCC for both HIV-1 andSIV envelope glycoproteins (Fig. 2A). In accordance with ADCC activity, PGT145 effi-ciently stained HIV-1NL4-3-infected cells and displayed low but detectable binding toHIV-1JR-FL- and SHIVAD8-EO-infected cells (Fig. 2A). PGT145 staining of SIV-infected cellswas highest for cells infected with SIVmac316, followed by SIVmac239 and SIVsmE543-3(Fig. 2A). Moreover, Env staining with PGT145 correlated with ADCC activity(P � 0.0314, Pearson correlation) (Fig. 2B).

FIG 2 PGT145 stains cells infected with diverse lentiviral isolates. (A) Overlay plots show PGT145 (blue) versusDEN3 (red) staining of CEM.NKR-CCR5-sLTR-Luc cells infected with the indicated viruses. (B) Area-under-the-curve(AUC) values for ADCC responses were compared to the geometric mean fluorescence intensity (gMFI) of PGT145staining on the surface of virus-infected cells by Pearson correlation. Virus-infected cells were identified by gatingon the Gag� CD4low population, and PGT145 staining was detected with PE-conjugated anti-human IgG F(ab=)2.

PGT145 Directs ADCC against SIVsmm/mac-Infected Cells ®

July/August 2019 Volume 10 Issue 4 e01255-19 mbio.asm.org 3

on May 25, 2020 by guest

http://mbio.asm

.org/D

ownloaded from

Since species-specific differences in the glycosylation of Env may affect the bindingof antibodies to glycan-dependent epitopes, we wanted to confirm that the staining ofSIV-infected cells by PGT145 was not an artifact of SIV Env glycosylation in human cells.PGT145 was therefore tested for binding to SIV-infected rhesus macaque CD4� lym-phocytes. Activated PBMCs were infected with wild-type SIVmac239 or SIVmac239 Y721G,which contains a tyrosine-to-glycine substitution in a conserved endocytosis motif ofthe gp41 cytoplasmic domain that increases Env levels on the surface of infected cells(23, 24), and PGT145 staining was assessed by flow cytometry. In contrast to the controlantibody (DEN3), PGT145 stained over 90% of the SIV-infected (CD4low Gag�) cells(Fig. 3A). Consistent with Env-dependent binding, the fluorescence intensity of PGT145staining was 1.6-fold higher on cells infected with SIVmac239 Y721G than on cellsinfected with wild-type SIVmac239 (Fig. 3A). PGT145 is therefore able to recognize SIVEnv expressed on the surface of primary rhesus macaque lymphocytes in addition tohuman CD4� T cell lines.

PGT145, which is a human IgG1 antibody, can also be recognized efficiently byrhesus macaque CD16. ADCC responses to SIV-infected cells were observed using NKcell lines expressing either human or rhesus macaque CD16 (Fig. 3B). Moreover, primaryNK cells from two different rhesus macaques mediated similar ADCC responses in thepresence of PGT145 (Fig. 3C). Thus, in addition to binding to Env on SIV-infectedprimary CD4� T cells, PGT145 can interact with Fc� receptors on the surface ofmacaque lymphocytes to mediate ADCC responses.

PGT145 neutralizes HIV-1 and SHIV, but not SIVsmm/mac. Since ADCC generallycorrelates with neutralization (12–14), we tested PGT145 for the ability to neutralize thesame viruses that were susceptible to ADCC. As expected, PGT145 potently neutralizedviruses expressing HIV-1 Env proteins, including HIV-1NL4-3, HIV-1JR-FL, and SHIVAD8-EO

FIG 3 PGT145 binds to SIV-infected primary rhesus macaque CD4� T cells and is recognized by rhesusmacaque CD16. (A) Overlay plots show PGT145 (blue) versus DEN3 (red) staining of activated rhesusmacaque CD4� T cells infected with wild-type SIVmac239 or SIVmac239 Y721G. Virus-infected cells wereidentified by gating on the Gag� CD4low population, and PGT145 staining was detected with PE-conjugated anti-human IgG F(ab=)2. (B and C) ADCC responses were measured using an NK cell lineexpressing either human or rhesus macaque CD16 (B) or unstimulated PBMCs from two differentmacaques (C) by incubating SIV-infected CEM.NKR-CCR5-sLTR-Luc cells at a 10:1 effector-to-target-cellratio in the presence of the indicated concentrations of PGT145. The dotted line indicates half-maximalkilling, and the error bars represent standard deviation of the mean from triplicate wells.

von Bredow et al. ®

July/August 2019 Volume 10 Issue 4 e01255-19 mbio.asm.org 4

on May 25, 2020 by guest

http://mbio.asm

.org/D

ownloaded from

(Fig. 4A). Corresponding to their sensitivity to ADCC, neutralization was most potent forHIV-1NL4-3 (IC50 � 0.004 �g/ml), followed by SHIVAD8-EO (IC50 � 0.28 �g/ml) andHIV-1JR-FL (IC50 � 0.62 �g/ml) (Fig. 4A). In contrast, despite being able to bind to Env onthe surface of SIV-infected cells, PGT145 failed to neutralize any of the SIV isolates(Fig. 4B).

PGT145 binds to a conserved conformational epitope at the V2 apex. Thebinding of PGT145 to HIV-1 Env requires an N-linked glycan at position 160 and a patchof basic amino acids in the V2 apex of the envelope trimer (16, 17). To determine ifPGT145 recognizes similar features of SIVmac239 Env, we introduced substitutions intothe corresponding region of the SIVmac239 V2 loop and tested these mutants forsusceptibility to ADCC (Fig. 5A). Elimination of a predicted glycosylation site at position171 (N171Q), corresponding to N160 in HIV-1HXB2 Env, abolished the ability of PGT145to mediate the lysis of infected cells (Fig. 5B and C), indicating that the recognition ofSIVmac239 Env by this antibody is dependent on an N-linked glycan at this position. Anarginine-to-alanine substitution at position 177 (R177A), corresponding to R166 inHIV-1HXB2, also greatly diminished the susceptibility of virus-infected cells to ADCC(Fig. 5B and C). Whereas lysine-to-serine substitutions at positions 179 and 181 had littleeffect (Fig. 5B and C), cells infected with a virus containing a substitution at position 180(K180S) showed increased sensitivity to ADCC (Fig. 5B and C). Remarkably, this singleamino acid change also rendered the virus susceptible to PGT145 neutralization(Fig. 5D).

Next, we investigated the relationship between ADCC and infected cell binding forthe Env variants by flow cytometry. Whereas little or no staining was observed for cellsinfected with the N171Q and R177A variants at antibody concentrations as high as50 �g/ml, PGT145 staining was detectable on cells infected with the K180S variant at3.2 ng/ml, which was approximately 1,000-fold lower than the concentration necessaryfor a similar level of Env staining on cells infected with wild-type SIVmac239 (Fig. 5E).Thus, the avidity of PGT145 binding to Env on cells infected with the K180S variant ismuch higher than on cells infected with wild-type SIV. PGT145 binding to Env alsocorrelated strongly with ADCC responses to the V2 variants (Fig. 5F) (P � 0.0001,Pearson correlation); however, staining with SIV� plasma did not differ (Fig. 5G),indicating that differences in PGT145 binding reflect specific interactions with the V2apex, rather than nonspecific differences in Env expression.

PGT145 binding to SIV Env trimers. Given the correlation between PGT145binding to SIV-infected cells and ADCC, we hypothesized that differences in thesusceptibility of the V2 variants to ADCC are a function of differences in the binding ofPGT145 to Env trimers. To investigate this possibility, we generated soluble SIVmac239Env SOSIP.664 trimers with substitutions in V2 that increase or decrease sensitivity toADCC and analyzed their binding kinetics to PGT145 by biolayer interferometry (BLI).Whereas PGT145 binding varied (Fig. 6A), CD4-IgG2 bound to each of the SOSIP trimers

FIG 4 PGT145 neutralizes HIV-1 and SHIV but not SIV. The indicated viruses were incubated with serial dilutionsof PGT145 for 1 h at 37°C before addition to TZM-bl cells (A) or C8166-SEAP cells (B). After 3 days, neutralization wasmeasured as a decrease in luciferase (A) or secreted alkaline phosphatase (SEAP) (B) activity (RLU) compared tomock-infected or virus-infected control wells without antibody. The dotted line indicates 50% neutralization.

PGT145 Directs ADCC against SIVsmm/mac-Infected Cells ®

July/August 2019 Volume 10 Issue 4 e01255-19 mbio.asm.org 5

on May 25, 2020 by guest

http://mbio.asm

.org/D

ownloaded from

FIG 5 PGT145 binds to the V2 apex of SIVmac239 Env. (A) Amino acid substitutions were introduced in the V2region of SIVmac239 Env. The amino acid sequence of HIV-1HXB2 is shown for reference, and the positions indicatedin the legend correspond to SIVmac239 Env. Dots indicate identity, and dashes indicate gaps. (B) CEM.NKR-CCR5-sLTR-Luc cells were infected with SIVmac239 carrying the indicated Env mutations and incubated with an NK cellline expressing human CD16 in the presence of serial dilutions of PGT145. ADCC activity was measured as thedose-dependent loss of luciferase activity in % RLU. The dotted line represents half-maximal lysis of infected cells,and the error bars represent the standard deviation of the mean from triplicate wells. (C) Area-under-the-curve(AUC) values for ADCC were calculated and compared by one-way ANOVA with a Holm-Sidak adjustment formultiple comparisons. (D) Wild-type SIVmac239 and SIVmac239 K180S were incubated with serial dilutions of PGT145for an hour before infecting C8166-SEAP cells. Neutralization was determined by calculating the loss of secretedalkaline phosphatase (SEAP) activity. The dotted line indicates 50% neutralization, and the error bars represent thestandard deviation of the mean from triplicate wells. (E to G) CEM.NKR-CCR5-sLTR-Luc cells were infected withSIVmac239 V2 variants and stained with either PGT145 (E) or plasma from an SIVmac239-infected rhesus macaque (G).Fluorescence intensities are shown for viable infected (Gag� CD4low) cells. PGT145 staining (AUC) was comparedto ADCC (AUC) by Pearson correlation (F).

von Bredow et al. ®

July/August 2019 Volume 10 Issue 4 e01255-19 mbio.asm.org 6

on May 25, 2020 by guest

http://mbio.asm

.org/D

ownloaded from

with similar kinetics (Fig. 6B). As expected, binding was not detectable for the HIV-1CD4 binding site antibody VRC01 or the dengue virus-specific control antibody DEN3(Fig. 6C and D). Consistent with the specificity of PGT145 for Env trimers, cryo-electronmicroscopy 2-D class averages confirm that SIV Env affinity purified with PGT145 formswell-folded trimers (see Fig. S1 in the supplemental material). BLI analysis revealed thatPGT145 bound to the wild-type Env trimer with a dissociation constant (Kd) of 44 nM(Fig. 6A and E). Similar to the effects of the N171Q and R177A substitutions on PGT145binding to infected cells and ADCC responses, these changes eliminated or dramaticallyreduced PGT145 binding to SOSIP trimers (Fig. 6A). Conversely, the K180S substitutionincreased the trimer binding affinity of PGT145 (Fig. 6A and E), primarily as a functionof a lower off-rate (Fig. 6A and F). There were, however, some differences in PGT145binding to recombinant Env trimers versus infected cells. Compared to infected cellbinding, the range of PGT145 affinities for the SOSIP trimer variants was more com-pressed and the rank order exhibited some differences. The explanation for thesedifferences is presently unclear but may reflect differences in the valency of PGT145binding to Env on the surface of virus-infected cells versus recombinant SOSIP trimersor subtle differences in Env conformation and/or V2 glycosylation on infected cellscompared to SOSIP trimers.

Taken together, these results suggest that differences in the sensitivity of the SIV V2variants to ADCC versus neutralization can be explained by differences in the affinity of

FIG 6 PGT145 binding to SIV Env trimers with V2 apex substitutions. The binding kinetics of PGT145 (A), CD4-IgG2(B), VRC01 (C), and DEN3 (D) to soluble SIVmac239 SOSIP.664 Env trimers with the indicated substitutions in the V2region was assessed by biolayer interferometry. Monoclonal antibodies immobilized on anti-human IgG-Fc-coatedbiosensors were dipped into trimer solutions (500 nM). The MAb-trimer interaction is represented as binding curvesshowing the rates of association (180 to 300 s) and dissociation (300 to 540 s). The dissociation constants (Kd) (E)and dissociation rates (Koff) (F) are plotted for PGT145 binding to each of the Env trimers.

PGT145 Directs ADCC against SIVsmm/mac-Infected Cells ®

July/August 2019 Volume 10 Issue 4 e01255-19 mbio.asm.org 7

on May 25, 2020 by guest

http://mbio.asm

.org/D

ownloaded from

PGT145 binding to Env trimers. Whereas the binding affinity of PGT145 to the wild-typeSIV Env trimer appears to be sufficient to trigger ADCC responses by linking Env spikeson the surface of virus-infected cells to Fc receptors on NK cells, it is not of sufficientaffinity for Env trimers on virions to permit neutralization of viral infectivity. In the caseof the K180S variant, this single amino acid substitution raises the binding affinity byreducing the dissociation rate of PGT145-Env trimer complexes to an extent thatSIV-infected cells become more sensitive to ADCC and cell-free virus becomes suscep-tible to neutralization.

DISCUSSION

Owing to the independent evolutionary histories of HIV-1 and SIVsmm/mac and to therapid evolution of their viral envelope glycoproteins in response to host immuneresponses, antibodies to the HIV-1 Env protein, including broadly neutralizing antibod-ies (bNAbs), typically do not cross-react with the Env proteins of SIVsmm/mac. In thepresent study, we show that the HIV-1 Env-specific bNAb PGT145 can opsonizeSIV-infected cells for elimination by ADCC but does not neutralize SIV infectivity. Thiscross-reactivity is not strain specific, since PGT145 mediated ADCC against cells infectedwith independent SIV isolates, nor is it a result of differences in Env glycosylation inhuman cells, since PGT145 also bound to Env expressed on the surface of SIV-infectedprimary rhesus macaque lymphocytes. The recognition of SIV Env by PGT145 wasfurther verified by showing that substitutions in V2 corresponding to the HIV-1 V2epitope for this antibody modulate the sensitivity of SIV-infected cells to ADCC.

A recent structure revealed that the CDRH3 loop of PGT145 forms a long �-hairpinthat reaches down through the 3-fold axis of symmetry of the Env trimer to makecontact with all three gp120 protomers (17). PGT145 binding to the Env trimer ismonomeric and depends on asymmetric contacts with an N-linked glycan at position160 (N160) and on electrostatic interactions between sulfotyrosine residues at the tipof the CDRH3 loop and basic residues of the V2 core (17). Although the precisemolecular interactions between PGT145 and SIV Env are less clear, the dependence ofPGT145 on SIV V2 residues N171 and R177, which correspond to HIV-1 V2 residues N160and R166, respectively, indicates that PGT145 probably binds to SIV Env in a very similarfashion. In accordance with the contribution of residues N160 and R166 to the PGT145epitope in HIV-1 Env, N171Q and R177A substitutions in the V2 loop of SIV dramaticallyreduce PGT145 binding and the sensitivity of SIV-infected cells to ADCC. Thus, PGT145recognizes a conserved conformational epitope at the V2 apex of the SIV Env trimer.

The ability of PGT145 to mediate ADCC without neutralizing viral infectivity can beexplained by its relatively low affinity for the SIV V2 apex. Sensitivity to ADCC stronglycorrelated with PGT145 binding to Env on the surface of SIV-infected cells, andalthough PGT145 did not block the infectivity of wild-type SIV, a single lysine-to-serinechange at position 180 (K180S) was sufficient to render the virus susceptible toneutralization. Consistent with the effect of this change on increasing PGT145 bindingto Env on infected cells, the K180S substitution increased the affinity of PGT145 forsoluble SIV Env SOSIP.664 trimers as measured by biolayer interferometry. The higheraffinity of PGT145 for the K180S trimer was primarily a function of a lower dissociationrate. These results suggest that the affinity of PGT145 binding to wild-type SIV Env,while sufficient to trigger ADCC by cross-linking multiple Env trimers on SIV-infectedcells to Fc receptors on NK cells, is not high enough to occupy enough Env trimers onvirions to block SIV infectivity. However, by stabilizing PGT145 binding to Env, theK180S substitution increases the affinity of PGT145 for Env on virions to a level sufficientto make the virus susceptible to neutralization.

PGT145 has a unique germ line configuration and an extended CDRH3 loop thatdistinguishes it from other V2 apex bNAbs (15, 16, 25). Accordingly, other V2 apex-specific bNAbs such as PG9 and PG16, which bind to similar quaternary epitopesconsisting of an N-linked glycan and underlying basic residues, do not recognizeSIVsmm/mac-infected cells (12). Thus, the cross-reactivity of PGT145 with SIVsmm/mac Envappears to reflect unique structural features of this antibody. However, PGT145, as well

von Bredow et al. ®

July/August 2019 Volume 10 Issue 4 e01255-19 mbio.asm.org 8

on May 25, 2020 by guest

http://mbio.asm

.org/D

ownloaded from

as PG9 and PG16, is able to neutralize a subset of SIVcpzPtt isolates (26). Although SIVcpz

is much more closely related to HIV-1 than SIVsmm/mac, of the different classes of bNAbsthat were tested, only 4E10 and 10E8, which target a well-conserved epitope at themembrane-proximal external region (MPER) of gp41, exhibited a similar breadth ofneutralization (26). These observations therefore suggest a surprising degree of con-servation of the V2 apex among phylogenetically diverse primate lentiviruses.

Conservation of the V2 apex may reflect important functional constraints on thisregion of Env. In order for Env to mediate fusion of the viral and cellular membranes,the trimer must transition from a closed to an open conformation in response to CD4and coreceptor engagement that ultimately exposes the gp41 fusion peptide. Prior toCD4 binding, the Env trimer exists in a closed, metastable conformation held togetherlargely by hydrophobic interactions (17). The proximity of positively charged residuesat the core of the V2 apex as a result of the coalescence of basic residues from adjacentgp120 subunits may provide a destabilizing force that is necessary for the trimer totransition to an open conformation (17). Thus, while surface-exposed residues and theposition of N-linked glycans may vary, the core of the V2 apex is more highly conservedto preserve key features essential for Env function. Consistent with this idea, structuraldata suggest that the breadth of HIV-1 neutralization by PGT145 is determined in partby general electrostatic interactions between anionic residues at the tip of the CDRH3loop and an electropositive sink formed by basic residues at the core of the V2 apex,with relatively few contacts with specific residues (17).

Although instances of neutralization in the absence of ADCC have been observed,most antibodies that mediate ADCC against HIV-infected cells also neutralize viralinfectivity (12–14). Thus, the ability of PGT145 to direct NK cell killing of SIV-infectedcells without blocking SIV infectivity is unusual. In this case, the uncoupling of ADCCfrom neutralization can be explained by the relatively low affinity of PGT145 for the V2apex of SIV Env, rather than differences in epitope exposure on the surface of infectedcells versus virions. These findings illustrate how quantitative differences in antibodybinding to Env can result in qualitative differences in antiviral activity with importantimplications for the selection of antibodies as immunotherapies to deplete HIV-1reservoirs and for the development of antibody-based vaccines.

MATERIALS AND METHODSViruses. Substitutions in SIVmac239 Env were introduced in the hemiviral plasmid p239SpE3= by

site-directed mutagenesis. SphI-PmlI fragments carrying the mutations were subcloned into the full-length infectious molecular clone SIVmac239 SpX. The infectious molecular clone for SIVsmE543-3 (22) wasprovided by Vanessa Hirsch, National Institute of Allergy and Infectious Diseases, National Institutes ofHealth.

Virus production. Virus stocks were produced by transfection of proviral DNA into HEK293T cellsusing GenJet transfection reagent (SignaGen). Culture supernatants were collected 48 and 72 h post-transfection, cell debris was removed by centrifugation, and aliquots of virus-containing supernatantwere stored at �80°C. Virus concentrations were determined by anti-p27 or anti-p24 enzyme-linkedimmunosorbent assay (ABL, Inc.).

ADCC assay. ADCC activity was measured as previously described (27, 28). Target cells expressingluciferase (Luc) upon infection (CEM.NKR-CCR5-sLTR-Luc) were infected by spinoculation in the presenceof 40 �g/ml Polybrene. At 4 days postinfection, cells were incubated with serial dilutions of PGT145 andan NK cell line (KHYG-1 cells) expressing either human or rhesus macaque CD16 at a 10:1 effector-to-target-cell ratio. For assays with primary cells, SIV-infected target cells were incubated with unstimulatedrhesus macaque PBMCs. The dose-dependent loss of Luc activity after 8 h was measured as an indicationof antibody-mediated killing of virus-infected cells. Infected target cells incubated with NK cells in theabsence of antibody were used to measure maximal Luc activity, and uninfected target cells culturedwith NK cells were used to determine background Luc activity.

Neutralization assay. Neutralization of viral infectivity was measured using reporter cell assays aspreviously described (9, 29, 30). Neutralization of SIV isolates was tested using C8166 cells expressingsecreted alkaline phosphatase (SEAP) under the control of the SIV LTR (29). SIVmac239 (0.5 ng p27),SIVmac316TM-open (15 ng p27), or SIVsmE543-3 (2 ng p27) was incubated with serial dilutions of antibodyfor 1 h at 37°C before adding 15,000 C8166-SEAP cells per well. After 3 days, SEAP activity was measuredusing a Phospha-Light SEAP detection kit (Applied Biosystems), and virus neutralization was calculatedfrom reductions in RLU relative to cells incubated with virus but no antibody. Uninfected cells weremeasured to account for background SEAP activity. Neutralization of HIV-1 and SHIV isolates was similarlytested using a standard reporter assay (9, 30). After a 1-h incubation at 37°C with different concentrationsof PGT145, 4 ng p24 (HIV-1NL4-3), 10 ng p24 (HIV-1JR-FL), or 20 ng p27 (SHIVAD8-EO) per well was used to

PGT145 Directs ADCC against SIVsmm/mac-Infected Cells ®

July/August 2019 Volume 10 Issue 4 e01255-19 mbio.asm.org 9

on May 25, 2020 by guest

http://mbio.asm

.org/D

ownloaded from

infect TZM-bl luciferase reporter cells seeded at 5,000 cells per well the previous day. Neutralization wasdetermined by measuring the loss of luciferase activity relative to cells infected in the absence ofantibody, after deducting background luciferase activity from uninfected cells.

Primary cells. Peripheral blood mononuclear cells were isolated from whole blood using Ficoll-Pacque Plus (GE Healthcare). CD4� T cells were subsequently enriched using the EasySep human CD4�

T cell isolation kit (StemCell Technologies) and activated using Dynabeads human T activator CD3/CD28microbeads (Gibco). Activated CD4� T cells were cultured in RPMI 1640 (HyClone) supplemented with20% heat-inactivated FBS (HyClone), 2 mM L-glutamine (HyClone), 100 �g/ml Primocin (Invivogen), and30 U/ml interleukin-2 (provided by Maurice Gately, Hoffmann-La Roche Inc., through the NIH AIDSReagent Program). Infections were performed 3 days postactivation by spinoculation in the presence of40 �g/ml Polybrene (Millipore).

Flow cytometry. Surface staining for Env was performed 3 days postinfection as previously de-scribed (23, 31). Antibody binding to Env was detected using a monoclonal antibody followed by a PE-or AF647-conjugated polyclonal anti-human IgG F(ab=)2. Cells were surface stained for CD45 (peridininchlorophyll protein [PerCP]; clone 2D1) and CD4 (Alexa Fluor 700; clone RPA-T4) and then permeabilizedand stained for intracellular Gag (fluorescein isothiocyanate [FITC]; clone 55-2F12). Nonviable cells wereexcluded using Live/Dead fixable dead cell aqua stain (Invitrogen), and data were collected using a SORPBD LSR-II flow cytometer (Becton Dickinson). After gating on viable CD45� CD4low Gag� cells, thegeometric mean fluorescence intensity (gMFI) of Env staining was calculated using FlowJo, version 9.7.7(Tree Star, Inc.).

Expression and purification of soluble SIVmac239 trimer. SOSIP.664 Env trimer modifications wereincorporated into the SIVmac239 Env coding sequence, as described previously (32). Amino acid substi-tutions were incorporated by using a QuikChange site-directed mutagenesis kit (Agilent Technologies,USA), according to the manufacturer’s instructions. All the mutations were confirmed by DNA sequenceanalysis (Eton Bioscience, San Diego, CA). Soluble recombinant Env trimers were expressed in HEK293Fcells as described elsewhere (32). Briefly, plasmids encoding the SIVmac239 SOSIP.664 trimer and its V2variants were cotransfected with a furin expression construct into HEK293F cells at a 3:1 ratio usingPEI-MAX 4000 transfection reagent (Polysciences, Inc.). The secreted trimer proteins were purified fromcell supernatants after 5 days using agarose-bound Galanthus nivalis lectin (GNL) (Vector Labs) columnsas described previously (32). Affinity-purified proteins were separated by size exclusion chromatographyusing Superdex 200 10/300 GL columns (GE Healthcare) in phosphate-buffered saline (PBS). The purifiedtrimers were stored at �80°C until use.

Biolayer interferometry binding assay. The binding kinetics of PGT145 and control MAbs with theaffinity-purified trimers were analyzed by biolayer interferometry (BLI) using an Octet K2 system(FortéBio; Pall Life Sciences), as described previously (33). Briefly, the MAbs at concentrations of 10 �g/mlin PBS with 0.1% Tween (PBST) were immobilized onto anti-human IgG-Fc biosensors (AHC; FortéBio) for60 s to achieve a binding response of at least 1.0 nm. After washing away the unbound MAb, theMAb-immobilized biosensor was dipped into a solution containing SOSIP.664 trimer protein (at a finalconcentration of 500 nM) as analyte and incubated for 120 s at 1,000 rpm. This was done by washing offthe unbound trimer by placement into the PBST buffer for 240 s at 1,000 rpm. The 120-s and 240-sbinding intervals denote the association and dissociation curves, respectively. The kinetic fits (1:1 bindingkinetics model) were performed with the FortéBio Data Analysis version.9 software using the globalfitting function.

Cryo-electron microscopy. SIVmac239 Env trimers were affinity purified with PGT145 and analyzedby cryo-electron microscopy. Three microliters of �6-mg/ml Env trimer was diluted 3:1 into 1.8 mmn-dodecyl-�-D-maltoside and plunge frozen on 1.2/1.3 C-Flat holey carbon grids (Protochips) with an FEIVitrobot Mark IV (Thermo Fisher). Data were collected on an FEI Titan Krios instrument (Thermo Fisher)operating at 300 keV at a defocus range of – 0.6 to �2.6 �m and controlled using Leginon (34). Moviemicrographs were captured on a K2 Summit direct electron detector (Gatan) using an electron dose rateof �4.65 e�/pixel/s, a 250 ms�1 frame rate, and a total exposure time of 11.5 s, for a total dose of �50.4e�/Å2. Frames were aligned and dose weighted with MotionCor2 (35), and the contrast-transfer functionswere estimated with Gctf (36). The final magnified pixel size was 1.03 Å. Particles were picked andextracted with Relion2 automated particle picking using a Gaussian disc as a template (37). 2-Dclassification was performed using CryoSparc2 (38). The final set of 2-D classes shown in Fig. S1 in thesupplemental material is composed of �62,000 particles.

SUPPLEMENTAL MATERIALSupplemental material for this article may be found at https://doi.org/10.1128/mBio

.01255-19.FIG S1, TIF file, 1.4 MB.

ACKNOWLEDGMENTSThis work was supported by the National Institutes of Health grants AI121135,

AI095098, AI098485, AI055332, UM1 AI100663, and OD011106. D.T.E. is an ElizabethGlaser Scientist of the Elizabeth Glaser Pediatric AIDS Foundation.

von Bredow et al. ®

July/August 2019 Volume 10 Issue 4 e01255-19 mbio.asm.org 10

on May 25, 2020 by guest

http://mbio.asm

.org/D

ownloaded from

REFERENCES1. Hahn BH, Shaw GM, Cock KMD, Sharp PM. 2000. AIDS as a zoonosis:

scientific and public health implications. Science 287:607– 614. https://doi.org/10.1126/science.287.5453.607.

2. Gao F, Bailes E, Robertson DL, Chen Y, Rodenburg CM, Michael SF,Cummins LB, Arthur LO, Peeters M, Shaw GM, Sharp PM, Hahn BH. 1999.Origin of HIV-1 in the chimpanzee Pan troglodytes. Nature 397:436 – 441.https://doi.org/10.1038/17130.

3. Keele BF, Jones JH, Terio KA, Estes JD, Rudicell RS, Wilson ML, Li Y, LearnGH, Beasley TM, Schumacher-Stankey J, Wroblewski E, Mosser A, RaphaelJ, Kamenya S, Lonsdorf EV, Travis DA, Mlengeya T, Kinsel MJ, Else JG,Silvestri G, Goodall J, Sharp PM, Shaw GM, Pusey AE, Hahn BH. 2009.Increased mortality and AIDS-like immunopathology in wild chimpan-zees infected with SIVcpz. Nature 460:515–519. https://doi.org/10.1038/nature08200.

4. Hirsch VM, Olmsted RA, Murphey-Corb M, Purcell RH, Johnson PR. 1989.An African primate lentivirus (SIVsm) closely related to HIV-2. Nature339:389 –392. https://doi.org/10.1038/339389a0.

5. Apetrei C, Kaur A, Lerche NW, Metzger M, Pandrea I, Hardcastle J,Falkenstein S, Bohm R, Koehler J, Traina-Dorge V, Williams T, Staprans S,Plauche G, Veazey RS, McClure H, Lackner AA, Gormus B, RobertsonDL, Marx PA. 2005. Molecular epidemiology of simian immunodeficiencyvirus SIVsm in U.S. primate centers unravels the origin of SIVmac andSIVstm. J Virol 79:8991–9005. https://doi.org/10.1128/JVI.79.14.8991-9005.2005.

6. Apetrei C, Lerche NW, Pandrea I, Gormus B, Silvestri G, Kaur A, RobertsonDL, Hardcastle J, Lackner AA, Marx PA. 2006. Kuru experiments triggeredthe emergence of pathogenic SIVmac. AIDS 20:317–321. https://doi.org/10.1097/01.aids.0000206498.71041.0e.

7. Jurkiewicz E, Hunsmann G, Hayami M, Ohta Y, Schmitz H, Schneider J.1988. Serological and structural comparison of immunodeficiency vi-ruses from man, African green monkey, rhesus monkey and sootymangabey. Z Naturforsch C 43:449 – 454. https://doi.org/10.1515/znc-1988-5-621.

8. Sutjipto S, Kodama T, Yee J, Gettie A, Jennings M, Desrosiers RC, Marx PA.1990. Characterization of monoclonal antibodies that distinguish simianimmunodeficiency virus isolates from each other and from human im-munodeficiency virus types 1 and 2. J Gen Virol 71:247–249. https://doi.org/10.1099/0022-1317-71-1-247.

9. Wei X, Decker JM, Wang S, Hui H, Kappes JC, Wu X, Salazar-Gonzalez JF,Salazar MG, Kilby JM, Saag MS, Komarova NL, Nowak MA, Hahn BH,Kwong PD, Shaw GM. 2003. Antibody neutralization and escape byHIV-1. Nature 422:307–312. https://doi.org/10.1038/nature01470.

10. Richman DD, Wrin T, Little SJ, Petropoulos CJ. 2003. Rapid evolutionof the neutralizing antibody response to HIV type 1 infection. ProcNatl Acad Sci U S A 1000:4144 – 4149. https://doi.org/10.1073/pnas.0630530100.

11. Hatziioannou T, Evans DT. 2012. Animal models for HIV/AIDS research.Nat Rev Microbiol 10:852– 867. https://doi.org/10.1038/nrmicro2911.

12. von Bredow B, Arias JF, Heyer LN, Moldt B, Le K, Robinson JE, Zolla-Pazner S, Burton DR, Evans DT. 2016. Comparison of antibody-dependent cell-mediated cytotoxicity and virus neutralization by HIV-1Env-specific monoclonal antibodies. J Virol 90:6127– 6139. https://doi.org/10.1128/JVI.00347-16.

13. Bruel T, Guivel-Benhassine F, Amraoui S, Malbec M, Richard L, Bourdic K,Donahue DA, Lorin V, Casartelli N, Noel N, Lambotte O, Mouquet H,Schwartz O. 2016. Elimination of HIV-1-infected cells by broadlyneutralizing antibodies. Nat Commun 7:10844. https://doi.org/10.1038/ncomms10844.

14. Ren Y, Korom M, Truong R, Chan D, Huang SH, Kovacs CC, Benko E, SafritJT, Lee J, Garban H, Apps R, Goldstein H, Lynch RM, Jones RB. 2018.Susceptibility to neutralization by broadly neutralizing antibodies gen-erally correlates with infected cell binding for a panel of clade B HIVreactivated from latent reservoirs. J Virol 92:e00895-18. https://doi.org/10.1128/JVI.00895-18.

15. Walker LM, Huber M, Doores KJ, Falkowska E, Pejchal R, Julien JP, WangSK, Ramos A, Chan-Hui PY, Moyle M, Mitcham JL, Hammond PW, OlsenOA, Phung P, Fling S, Wong CH, Phogat S, Wrin T, Simek MD, Protocol G.Principal Investigators, Koff WC, Wilson IA, Burton DR, Poignard P. 2011.Broad neutralization coverage of HIV by multiple highly potent antibod-ies. Nature 477:466 – 470. https://doi.org/10.1038/nature10373.

16. Andrabi R, Voss JE, Liang CH, Briney B, McCoy LE, Wu CY, Wong CH,

Poignard P, Burton DR. 2015. Identification of common features inprototype broadly neutralizing antibodies to HIV envelope V2 apex tofacilitate vaccine design. Immunity 43:959 –973. https://doi.org/10.1016/j.immuni.2015.10.014.

17. Lee JH, Andrabi R, Su CY, Yasmeen A, Julien JP, Kong L, Wu NC, McBrideR, Sok D, Pauthner M, Cottrell CA, Nieusma T, Blattner C, Paulson JC,Klasse PJ, Wilson IA, Burton DR, Ward AB. 2017. A broadly neutralizingantibody targets the dynamic HIV envelope trimer apex via a long,rigidified, and anionic beta-hairpin structure. Immunity 46:690 –702.https://doi.org/10.1016/j.immuni.2017.03.017.

18. Daniel MD, Letvin NL, King NW, Kannagi M, Sehgal PK, Hunt RD, Kanki PJ,Essex M, Desrosiers RC. 1985. Isolation of T-cell tropic HTLV-III-like ret-rovirus from macaques. Science 228:1201–1204. https://doi.org/10.1126/science.3159089.

19. Letvin NL, Daniel MD, Sehgal PK, Desrosiers RC, Hunt RD, Waldron LM,Mackey JJ, Schmidt DK, Chalifoux LV, King NW. 1985. Induction ofAIDS-like disease in macaque monkeys with T-cell tropic retrovirusSTLV-III. Science 230:71–73. https://doi.org/10.1126/science.2412295.

20. Mori K, Rosenzweig M, Desrosiers RC. 2000. Mechanisms for adaptationof simian immunodeficiency virus to replication in alveolar macro-phages. J Virol 74:10852–10859. https://doi.org/10.1128/jvi.74.22.10852-10859.2000.

21. Mori K, Ringler DJ, Kodama T, Desrosiers RC. 1992. Complex determi-nants of macrophage tropism in env of simian immunodeficiency virus.J Virol 66:2067–2075.

22. Hirsch V, Adger-Johnson D, Campbell B, Goldstein S, Brown C, Elkins WR,Montefiori DC. 1997. A molecularly cloned, pathogenic, neutralization-resistant simian immunodeficiency virus, SIVsmE543-3. J Virol 71:1608 –1620.

23. von Bredow B, Arias JF, Heyer LN, Gardner MR, Farzan M, Rakasz EG,Evans DT. 2015. Envelope glycoprotein internalization protects humanand simian immunodeficiency virus infected cells from antibody-dependent cell-mediated cytotoxicity. J Virol 89:10648 –10655. https://doi.org/10.1128/JVI.01911-15.

24. Yuste E, Reeves J, Doms R, Desrosiers R. 2004. Modulation of Env contentin virions of simian immunodeficiency virus: correlation with cell surfaceexpression and virion infectivity. J Virol 78:6775– 6785. https://doi.org/10.1128/JVI.78.13.6775-6785.2004.

25. McLellan JS, Pancera M, Carrico C, Gorman J, Julien J-P, Khayat R, LouderR, Pejchal R, Sastry M, Dai K, O’Dell S, Patel N, Shahzad-Ul-Hussan S, YangY, Zhang B, Zhou T, Zhu J, Boyington JC, Chuang G-Y, Diwanji D,Georgiev I, Do Kwon Y, Lee D, Louder MK, Moquin S, Schmidt SD, YangZ-Y, Bonsignori M, Crump JA, Kapiga SH, Sam NE, Haynes BF, Burton DR,Koff WC, Walker LM, Phogat S, Wyatt R, Orwenyo J, Wang L-X, Arthos J,Bewley CA, Mascola JR, Nabel GJ, Schief WR, Ward AB, Wilson IA, KwongPD. 2011. Structure of HIV-1 gp120 V1/V2 domain with broadlyneutralizing antibody PG9. Nature 480:336 –343. https://doi.org/10.1038/nature10696.

26. Barbian HJ, Decker JM, Bibollet-Ruche F, Galimidi RP, West AP, Jr, LearnGH, Parrish NF, Iyer SS, Li Y, Pace CS, Song R, Huang Y, Denny TN,Mouquet H, Martin L, Acharya P, Zhang B, Kwong PD, Mascola JR, VerripsCT, Strokappe NM, Rutten L, McCoy LE, Weiss RA, Brown CS, Jackson R,Silvestri G, Connors M, Burton DR, Shaw GM, Nussenzweig MC, BjorkmanPJ, Ho DD, Farzan M, Hahn BH. 2015. Neutralization properties of simianimmunodeficiency viruses infecting chimpanzees and gorillas. mBio6:e00296-15. https://doi.org/10.1128/mBio.00296-15.

27. Alpert MD, Heyer LN, Williams DEJ, Harvey JD, Greenough T, Allhorn M,Evans DT. 2012. A novel assay for antibody-dependent cell-mediatedcytotoxicity against HIV-1- or SIV-infected cells reveals incomplete over-lap with antibodies measured by neutralization and binding assays. JVirol 86:12039 –12052. https://doi.org/10.1128/JVI.01650-12.

28. Alpert MD, Harvey JD, Lauer WA, Reeves RK, Piatak M, Jr, Carville A,Mansfield KG, Lifson JD, Li W, Desrosiers RC, Johnson RP, Evans DT. 2012.ADCC develops over time during persistent infection with live-attenuated SIV and is associated with complete protection againstSIVmac251 challenge. PLoS Pathog 8:e1002890. https://doi.org/10.1371/journal.ppat.1002890.

29. Means RE, Greenough T, Desrosiers RC. 1997. Neutralization sensitiv-ity of cell culture-passaged simian immunodeficiency virus. J Virol71:7895–7902.

30. Wei X, Decker JM, Liu H, Zhang Z, Arani RB, Kilby JM, Saag MS, Wu X,

PGT145 Directs ADCC against SIVsmm/mac-Infected Cells ®

July/August 2019 Volume 10 Issue 4 e01255-19 mbio.asm.org 11

on May 25, 2020 by guest

http://mbio.asm

.org/D

ownloaded from

Shaw GM, Kappes JC. 2002. Emergence of resistant human immunode-ficiency virus type 1 in patients receiving fusion inhibitor (T-20) mono-therapy. Antimicrob Agents Chemother 46:1896 –1905. https://doi.org/10.1128/aac.46.6.1896-1905.2002.

31. Arias JF, Heyer LN, von Bredow B, Weisgrau KL, Moldt B, Burton DR,Rakasz EG, Evans DT. 2014. Tetherin antagonism by Vpu protects HIV-infected cells from antibody-dependent cell-mediated cytotoxicity. ProcNatl Acad Sci U S A 111:6425– 6430. https://doi.org/10.1073/pnas.1321507111.

32. Sanders RW, Derking R, Cupo A, Julien JP, Yasmeen A, de Val N, Kim HJ,Blattner C, de la Pena AT, Korzun J, Golabek M, de Los Reyes K, Ketas TJ,van Gils MJ, King CR, Wilson IA, Ward AB, Klasse PJ, Moore JP. 2013. Anext-generation cleaved, soluble HIV-1 Env trimer, BG505 SOSIP.664gp140, expresses multiple epitopes for broadly neutralizing but notnon-neutralizing antibodies. PLoS Pathog 9:e1003618. https://doi.org/10.1371/journal.ppat.1003618.

33. Andrabi R, Su CY, Liang CH, Shivatare SS, Briney B, Voss JE, Nawazi SK,Wu CY, Wong CH, Burton DR. 2017. Glycans function as anchors forantibodies and help drive HIV broadly neutralizing antibody develop-

ment. Immunity 47:524 –537.e3. https://doi.org/10.1016/j.immuni.2017.08.006.

34. Potter CS, Chu H, Frey B, Green C, Kisseberth N, Madden TJ, MillerKL, Nahrstedt K, Pulokas J, Reilein A, Tcheng D, Weber D, Carragher B.1999. Leginon: a system for fully automated acquisition of 1000 electronmicrographs a day. Ultramicroscopy 77:153–161. https://doi.org/10.1016/S0304-3991(99)00043-1.

35. Zheng SQ, Palovcak E, Armache JP, Verba KA, Cheng Y, Agard DA. 2017.MotionCor2: anisotropic correction of beam-induced motion for im-proved cryo-electron microscopy. Nat Methods 14:331–332. https://doi.org/10.1038/nmeth.4193.

36. Zhang K. 2016. Gctf: real-time CTF determination and correction. J StructBiol 193:1–12. https://doi.org/10.1016/j.jsb.2015.11.003.

37. Scheres SH. 2015. Semi-automated selection of cryo-EM particles inRELION-1.3. J Struct Biol 189:114 –122. https://doi.org/10.1016/j.jsb.2014.11.010.

38. Punjani A, Rubinstein JL, Fleet DJ, Brubaker MA. 2017. cryoSPARC:algorithms for rapid unsupervised cryo-EM structure determination. NatMethods 14:290 –296. https://doi.org/10.1038/nmeth.4169.

von Bredow et al. ®

July/August 2019 Volume 10 Issue 4 e01255-19 mbio.asm.org 12

on May 25, 2020 by guest

http://mbio.asm

.org/D

ownloaded from